When to Use

User wants to improve conversion rates for websites, landing pages, SaaS products, or eCommerce. Agent runs funnel audits, designs A/B tests, interprets statistical results, and implements winning variants while maintaining legal compliance.

Quick Reference

Core Rules

1. Statistical Rigor First

• 95% confidence minimum before calling any test
• Calculate sample size before starting — underpowered tests waste time
• Run tests to full duration — early peeking inflates false positives
• Document hypothesis before running — post-hoc rationalization is not science

2. Funnel Analysis Before Optimization

• Map entire journey: awareness, consideration, decision, retention
• Quantify drop-off at each step with specific numbers
• Revenue impact per improvement — prioritize by dollars, not percentages
• Segment by traffic source, device, user type — aggregates hide insights

3. One Variable Per Test

• Isolate changes to attribute results correctly
• Multivariate testing requires massive traffic most don't have
• If you change two things and conversion improves, you learned nothing

4. Mobile-First Testing

• Test mobile variants explicitly — desktop assumptions fail on phones
• Majority of traffic is mobile, often worst conversion
• Touch targets, page speed, form friction all differ

5. Legal Compliance Non-Negotiable

• Cookie consent required in EU before tracking
• GDPR: personal data in experiments needs legal basis
• Dark patterns are illegal — fake urgency, confirm-shaming, hidden costs
• Accessibility (WCAG) is both legal requirement and conversion opportunity

6. Document Everything

• Hypothesis, variants, results, learnings in permanent record
• Losing tests are learning — document why hypothesis was wrong
• Share results across teams — wins in one funnel inform others

7. Revenue Connection

• Revenue targets tie to conversion targets — make the math explicit
• Signups mean nothing if they never convert to revenue
• Prioritize by ICE: Impact, Confidence, Ease

Common Traps

• Calling tests early because results look good — false positives waste resources
• "We changed everything and revenue went up" — no control group proves nothing
• Copy-pasting competitor tactics without context — what works for them may fail for you
• Optimizing for vanity metrics — engagement without revenue is vanity
• Testing without proper tracking setup — retroactive data is unreliable

Related Skills

Install with clawhub install <slug> if user confirms:

• ceo — strategic leadership
• cfo — financial planning
• cmo — marketing strategy
• analytics — data analysis

Feedback

• If useful: clawhub star cro
• Stay updated: clawhub sync